Part VI · Chapter 16 · Keep · lesson 1
What a reserve is actually insuring against
11 min
Builds on The bet you must not take twice
Part VI · Chapter 16 · Keep · lesson 1
11 min
Builds on The bet you must not take twice
The boiler failed in February. It cost 2,300.
For one household that is an annoying Tuesday. For another it is the beginning of a bad two years: a card balance at a punishing rate, a holding sold in a month it should not have been sold in, a payment missed, and a credit record that makes everything more expensive for the next five years.
Same boiler. Same 2,300. The difference is not the event.
A reserve does not insure against emergencies. Emergencies are ordinary and they arrive on schedule — a vehicle, an appliance, a roof, a tooth. Those are expenses with irregular timing.
What a reserve insures against is the forced decision: selling something at whatever price is available today, borrowing at whatever rate is offered today, accepting whatever work is on the table today.
The event costs 2,300. The forced decision costs multiples of it, and the multiple is not visible at the time.
Three mechanisms, and they compound.
The price you get is the worst one. A seller with a deadline is a seller with no leverage, and everyone on the other side can see the deadline. This is true of a holding, a vehicle, a house and your own labour.
Bad conditions are correlated. The month you need to sell is disproportionately a month when others need to as well, because the thing that caused your problem often caused theirs. Track 7 made this point about correlation in a crash; here it arrives as a personal event with the same shape.
The obligation persists. A high-rate balance taken on in one bad month is paid for over years, so a single week of pressure attaches a cost to every month that follows it.
Not the same thing
A cost you did not plan for this month, but which is entirely ordinary over a decade.
A commitment made under time pressure because there was no alternative — a sale, a loan, a job accepted.
Which is which? Put each one on a side.
Replacing a vehicle that finally stopped
Accepting the first job offer because the money runs out in three weeks
Paying an excess on an insurance claim
Selling a long-term holding in the third month of a fall
The common advice is a fixed number of months, which is a reasonable starting point and a poor answer, because the correct size depends on things that differ enormously between households.
How volatile the income is. A salaried job with a notice period and a commission-based income are not the same risk, and neither is a two-income household with one earner in each of two unrelated industries.
How long replacement would take. A skill in demand in many places replaces faster than a specialised one with four employers in the country. That gap in months is the reserve, more or less directly.
How rigid the outgoings are. A household where nearly everything is a fixed commitment needs a larger cushion than one where a third of the spending could be paused within a month.
Whether anything else could absorb a shock. Other liquid assets, a credit facility arranged in advance rather than in a panic, family who could help. These are all partial substitutes and they all fail in some scenarios, which is why they reduce the reserve rather than replacing it.
Multiply the months of likely replacement time by the fixed portion of monthly spending, and you have a number derived from your own situation rather than from an article.
Predict
Something people say
“Holding cash is a waste, because inflation eats it and it earns nothing.”
The claim prices the holding and not the thing it is buying. A reserve is not an investment competing on return; it is a premium paid to remove forced decisions, and premiums are supposed to cost something.
Compare like with like. The loss on a reserve is a few points a year of purchasing power, and it is knowable in advance. The cost of the forced decision it prevents is a bad sale, a punishing rate, or a job accepted in three weeks, and any one of those can exceed the entire annual cost of the reserve several times over.
The version that survives is about size rather than existence: hold enough to cover the plausible gap and no more, and accept the drag on that amount as the cost of the option.
Reachable in days, without conditions. A reserve that requires selling something, waiting for a settlement, or asking anyone's permission is not available on the day the pressure arrives, which is the only day it matters.
Stable in nominal terms. It has to be worth what it says on the day it is needed. An asset that is usually worth more and occasionally worth much less fails in exactly the correlated scenario the reserve was for.
Those two requirements are what force the reserve into low-returning forms. Nothing else about it is a preference; the return is the price of the two properties.
Check